OATML

Yi (Joshua) Ren is a postdoctoral researcher in the OATML Group at the University of Oxford, working under the supervision of Yarin Gal. His research takes a dynamics- and physics-inspired perspective on modern machine learning systems to understand how learning signals propagate through large neural networks and shape their representations and capabilities over time. He is particularly interested in modelling training as a dynamical system, where optimization induces forces, interactions, and constraints across different layers and features.

More broadly, his interests span large-scale representation learning, continual and lifelong learning, and the foundations of robustness and alignment in LLM-based systems. By combining theoretical analysis, physical analogies, and empirical investigation, his work seeks to build a more principled understanding of how complex learning systems evolve, and how they can be controlled more reliably in practice.

Before joining Oxford, Joshua completed his Ph.D. in Computer Science at the University of British Columbia in Canada, working on learning dynamics.